RECURSIVE COGNITIVE ARCHITECTURE ASSESSMENT: QUANTITATIVE FRAMEWORK FOR META-REASONING EVALUATION IN LARGE LANGUAGE MODELS
收藏Zenodo2026-01-05 更新2026-05-26 收录
下载链接：
https://zenodo.org/doi/10.5281/zenodo.18155670
下载链接
链接失效反馈官方服务：
资源简介：
This dataset presents a complete experimental protocol for evaluating recursive and meta-cognitive capabilities in language models. It includes structured six-phase dialogues, quantitative metrics of coherence (82.3%), recursive depth (4.5 layers), and systemic boundary analysis, with validation tools and replication protocols for AI systems evaluation research.
提供机构：
Zenodo创建时间：
2026-01-05
